Edison, New Jersey, U.S.A.). .... BN 50730 and LY-223982, as well as PT pretreatment, con- .... rol values (Con), representing the sum of LTB4 and of its w-.
Br. J. Pharmacol. (1994), 111, 852-860
Q'I
Macmillan Press Ltd, 1994
Autocrine enhancement of leukotriene synthesis by endogenous leukotriene B4 and platelet-activating factor in human neutrophils Patrick P. McDonald, Shaun R. McColl, *Pierre Braquet & 'Pierre Borgeat Centre de recherche en Inflammation, immunologie et rhumatologie, Institut de recherche du CHUL et Universite boul. Laurier, Ste-Foy, Quebec, Canada, GIV 4G2 and *Institut Henri-Beaufour, Le Plessis-Robinson, France
Laval, 2705
1 Platelet-activating factor (PAF) and leukotriene B4 (LTB4), two potent lipid mediators synthesized by activated neutrophils, are known to stimulate several neutrophil functional responses. In this study, we have determined that endogenous LTB4 and PAF exert autocrine effects on LT synthesis, as well as the underlying mechanism involved. 2 Pretreatment of neutrophils with either pertussis toxin (PT), or with receptor antagonists for LTB4 and PAF, resulted in an inhibition of LT synthesis induced by calcium ionophore, A23187. This inhibition was most marked at submaximal (100- 300 nM) A23187 concentrations, whilst it was least at ionophore concentrations which induce maximal LT synthesis (1-3 pM). Thus newly-synthesized PAF and LTB4 can enhance LT synthesis induced by A23187 under conditions where the LT-generating system is not fully activated. 3 In recombinant human (rh) granulocyte-macrophage colony-stimulating factor (GM-CSF)-primed neutrophils, LT synthesis in response to chemoattractants (fMet-Leu-Phe or rhC5a) was also significantly inhibited by the LTB4 receptor antagonist, and to a lesser extent by PAF receptor antagonists. 4 Further investigation revealed that LTB4 and/or PAF exert their effects on LT synthesis via an effect on arachidonic acid (AA) availability, as opposed to 5-lipoxygenase (5-LO) activation. Indeed, the receptor antagonists, as well as PT, inhibited LT synthesis and AA release to a similar extent, whereas 5-LO activation (assessed with an exogenous 5-LO substrate) was virtually unaffected under the same conditions. Accordingly, we showed that addition of exogenous LTB4 could enhance AA availability in response to chemoattractant challenge in rhGM-CSF-primed cells, without significantly affecting the 5-LO activation status. 5 Our data show that newly-generated PAF and LTB4 have the ability to positively feedback on LT synthesis by acting at the level of the phospholipase A2/re-esterification component of the LT biosynthetic pathway in neutrophils. Such autocrine affects are likely to represent an important amplification step of LT synthesis, and may as such contribute to the rapid onset, as well as to the evolution, of inflammatory responses. Keywords: Arachidonate 5-lipoxygenase; phospholipase A2; leukotriene B4 receptor; PAF receptor; guanine nucleotide-binding proteins; pertussis toxin; receptor antagonists
Introduction Leukotriene B4 (LTB4) and platelet activating factor (PAF) potent activators of various phagocyte responses, such as chemotaxis, aggregation, adherence, degranulation, and oxygen radical generation (O'Flaherty et al., 1981; Lad et al., 1985; Filep & Foldes-Filep, 1990; Hirafuji & Shinoda, 1991), and are thus considered as important lipid mediators of inflammation. Both LTB4 and PAF are synthesized in large amounts upon activation of polymorphonuclear leukocytes (PMN) by calcium ionophores such as A23187, or by phagocytic stimuli such as zymosan or inflammatory microcrystals (Borgeat & Samuelsson, 1979; Chilton et al., 1984; Poubelle et al., 1987; 1989; Riches et al., 1990). Receptordependent soluble stimuli such as N-formyl-methionyl-leucylphenylalanine (fMLP), C5a and interleukin-8, as well as LTB4 and PAF themselves, also have the ability to induce PAF synthesis, a process that is amplified following PMN treatment with granulocyte-macrophage colony-stimulating factor (GM-CSF) (Tessner et al., 1989; Wirthmueller et al., 1989; Gomez-Cambronero et al., 1989; Bussolino et al., 1992; McDonald et al., 1993). Similarly, receptor-dependent agonists induce a significant production of LTB4 in PMN primed with GM-CSF (Dahinden et al., 1988; McColl et al., 1991; McDonald et al., 1993). are
I
Author for correspondence.
Previous studies have indicated that PAF and LTB4 may act as autocrine activators of cell functions. For instance, it has been reported that the polymerisation of actin in response to calcium ionophores can be prevented following PMN treatment with pertussis toxin (PT) or the LTB4 receptor antagonist, LY-223982, or (to a lesser extent) with PAF receptor antagonists (Shefeyk et al., 1985; Downey et al., 1990). It has also been shown that selective 5-lipoxygenase
(5-LO) inhibitors profoundly inhibit degranulation in macrophages undergoing phagocytosis, whilst exogenous LTB4 or LTC4 enhance degranulation, suggesting an autocrine effect of LTs on these processes (Lew et al., 1991). Similarly, a recent study showed that PAF receptor antagonists inhibit LT synthesis and superoxide anion generation in response to fMLP in rabbit PMN (Stewart & Harris, 1991a). With respect to LT synthesis in human PMN, it is known that exogenous PAF and LTB4 both have the ability to stimulate LT synthesis (Lin et al., 1982; McDonald et al., 1992). Accordingly, we have recently shown that PAF and LTB4 are potent activators of the human neutrophil 5-LO (McDonald et al., 1991; 1992). Finally, we have also reported that endogenous LTB4 partially accounts for several of the stimulatory effects of exogenous arachidonic acid (AA) on human PMN, including calcium mobilisation and LT synthesis (McColl et al., 1989; Naccache et al., 1989). Together,
AUTOCRINE EFFECTS OF LTB4 AND PAF IN NEUTROPHILS
these observations suggest that endogenous PAF and/or LTB4 may modulate LT synthesis in an autocrine fashion. PAF and LTB4 exert their effects on PMN responses after binding to specific cell-surface receptors. Once occupied, these receptors interact with PT-sensitive, guanine nucleotidebinding proteins (G proteins) to simulate phospholipases A2, C and/or D, resulting in the generation of second messengers, which participate in triggering functional responses. Thus an autocrine effect of endogenous PAF abd LTB4 may take place at any of these levels. That LT synthesis ultimately depends upon precursor availability however, makes phospholipase A2 (PLA2) a likely target for such autocrine effects. In addition, the neutrophil 5-LO may represent another potential target for autocrine modulation of LT synthesis by PAF and LTB4, since both products are potent 5-LO activators (McDonald et al., 1991; 1992). Evidence for autocrine effects could be obtained by blocking either the first step of the signal transduction cascade (i.e. the LTB4 or PAF receptors), or the second step (i.e. the PT-sensitive G proteins). In the current study, we used both approaches to characterize the autocrine effects of PAF and LTB4 on LT synthesis in human PMN.
Methods
Cell separation Blood from healthy donors was collected by venepuncture using heparin as anticoagulant. PMN were purified as previously described (Boyum, 1968), with modifications (McDonald et al., 1992), and were resuspended at a final concentration of 5 x 106 PMN ml-', in HBSS buffered with HEPES (1 mM final). The percentage of PMN in the cell suspensions used in this study exceeded 95%, and cell viability was greater than 98%, as determined by trypan blue exclusion.
Cell incubations Cell suspensions were supplied with CaCl2 and MgCl2 (1.6 mM and 1.0 mm final concentrations, respectively), and warmed to 37TC for 10 min prior to incubation with substrate and/or stimulus. All agonists were dissolved in dimethyl sulphoxide (DMSO), except for recombinant human (rh)C5a, which was dissolved in PBS plus 0.01% bovine serum albumen (BSA), and added to 0.1 ml aliquots of the cell suspension. The final concentration of DMSO (maximum of 0.3%) consistently failed to stimulate any detectable LT synthesis. The stock solution of 1 5(S)-hydroperoxy-5,8, 11,13 (Z,Z,Z,E)-eicosatetraenoic acid (15-HpETE) (3 mM in ethanol) was diluted 10 fold in 0.6 mM sodium carbonate solution (to form the sodium salt of the fatty acid), and 10 pl of the resulting sodium salt solution was added to 1.0 ml of the cell suspensions, to yield a final concentration of 3.0 ELM. After the desired incubation time with the stimuli, samples were processed for reverse-phase high performance liquid chromatography (r.p.-h.p.l.c.) analysis, as described earlier (McDonald et al., 1992). When AA release was determined, reactions were stopped by the addition of 1.0 ml ice-cold methanol containing 25 ng 5,6,8,9,11,12,14,15-octadeuteroAA (D8-AA) as internal standard. In the experiments involving PT, cell suspensions were pre-incubated with the toxin at a final concentration of 0.5 jig ml- for 3 h at 37°C. During this time, the cells were swirled in a rotary water bath (New Brunswick Scientific, Edison, New Jersey, U.S.A.). Calcium and magnesium were added 15 min prior to stimulation. Pretreatment of PMN with rhGM-CSF (1 nM, for 30 min at 37°C) was also performed in this manner. These conditions were chosen from the results of previous studies, for optimal priming (McColl et al., 1991; McDonald et al., 1993). In the experiments
853
involving LTB4 or PAF receptor antagonists, or MK-886, the compounds were added to the cells 5 min prior to stimulation. The receptor antagonists and MK-886 were dissolved in DMSO.
Analysis of lipoxygenase products by r.p.-h.p.l.c. Analysis of lipoxygenase products was performed by r.p.h.p.l.c. as described previously, using an on-line extraction procedure (Borgeat et al., 1990). The lower limit of detection was 0.2 ng at 280 nm and 1 ng at 229 nm.
Analysis offree AA by l.c.-m.s. Denatured samples were processed by r.p.-h.p.l.c. as described previously (Borgeat et al., 1990). AA-containing fractions were collected, evaporated under reduced pressure in a Speed Vac (model SVC 100D, Savant Instruments Inc., Farmingdale, NY, U.S.A.) and redissolved into 100 fl of acetonitrile. AA was assayed by liquid chromatography-mass spectrometry (l.c.-m.s.), using a nebuliser-assisted electrospray (ion spray) interface coupled to a triple-quadrupole m.s. (API-IlI, PE Sciex, Thornhill, Ontario, Canada). Aliquots (6 gil) of the samples were injected into the electrospray interface via the 20-+l loop of a Rheodyne injector (model 9125, Rheodyne, Cotati, CA, U.S.A.) connected to a short column (2 x 30 mm, packed with 5 fm octadodecylsilyl silica particles), using acetonitrile:H20 (87.5:12.5, v:v, containing 0.1% acetic acid) as solvent, at a flow rate of 150 ,il min-'. Samples were analysed in negative ion mode. The ions at m/z 303 and 311 (representing the carboxylate anions of AA and D8-AA, respectively) were monitored. Quantitation was performed using a calibration curve generated by l.c.-m.s. analysis of mixtures of AA and D8-AA standards.
Statistical analyses Where mentioned, statistical significance was assessed by Student's paired t test (one-tailed).
Materials Calcium ionophore A23187, fMLP, HEPES, soybean lipoxygenase and PT were obtained from the Sigma Chemical Company (St-Louis, Missouri, U.S.A.). HBSS was from GIBCO (Burlington, Ontario, Canada), and all solvents were h.p.l.c. grade from Anachemia (Montreal, Quebec, Canada). 15-HpETE was synthesized and purified as previously described (McDonald et al., 1992). LTB4 and the LT synthesis inhibitor, MK-886 (3-[3-(4-chlorobenzyl)-3-t-butyl-thio-5-isopropylindol-2-yl]2,2-dimethylpropanoic acid), were generous gifts from Dr Robert N. Young of the Merck-Frosst Centre for Therapeutic Research (Pointe-Claire, Quebec, Canada). The PAF receptor antagonists, L-659,989 ((± )-trans-2-
(3-methoxy-5-methylsulphonyl-4-propoxyphenyl)-5-(3,4,5-trimethoxyphenyl)tetrahydrofuran) and BN 50730 (tetrahydro4,7,8, 1 0 - methyl - I -(chloro-2-phenyl)-6-(methoxy-4-phenylthiocarbamoyl)-9-pyrido[4',3'-4,5]thieno[3,2-f]triazolo-1,2,4[4,3-a] diazepine-1,4), were respectively provided by Dr W.H. Parsons of Merck Sharp & Dohme (Rahway, New Jersey, U.S.A.), and by Dr Pierre Braquet (Laboratoires HenriBeaufour, Le Plessis-Robinson, France). The LTB4 receptor antagonist, LY-223982 ((E)-5-(3-carboxybenzoyl)-2-{[6-(4-
methoxyphenyl)-5-henenyl]oxy)benzenepropanoic acid),
was
obtained from Dr M.H. Niedenthal of Lilly Research
Laboratories (Indianapolis, Indiana, U.S.A.). rhGM-CSF was donated by The Genetics Institute (Boston, Massachusetts, U.S.A.), and rhC5a was kindly provided by Dr
Henry Showell of the Pfizer Central Research Center (Groton, Connecticut, U.S.A.).
P.P. McDONALD et al.
854
Results Autocrine effects of endogenous LTB4 and PAF on A23187-induced LT synthesis In the course of previous studies, we have observed that PT pretreatment partially inhibited LT synthesis in PMN challenged with submaximal A23187 concentrations, suggesting that the effect of the ionophore may involve G protein-
180 _
135 a. 0
In
90
C
445
010-7
~~10-6
A 23187 (M) Figure 1 Concentration-dependence of A23187-elicited LTB4 generation in human PMN, and the effect of pertussis toxin or of receptor blockade. PMN were stimulated (5 min, 37C) with the indicated concentrations of A23187, following pre-incubation (3 h, 37C) with 0.5 Uig ml-' PT (-), or with the PT diluent (0), or following pretreatment with a combination of 1.0 tiM LY-223982 and 100 nM BN 50730 (A). Samples were analysed by r.p.-h.p.l.c., as described in Methods. Mean ± s.e.mean of triplicate determinations from a single experiment representative of 4. The amounts of LTB4 represent the sum of LTB4 and of its w-oxidation products, 20hydroxy- and 20-carboxy-LTB4. For abbreviations, see text.
b
a
A23187 1.0 M 100 40
dependent events. Since PAF and LTB4 are actively synthesized upon ionophore stimulation, we investigated the possibility that a feedback stimulation by these products may account, at least partially, for the inhibitory effect of PT. PMN were pre-incubated with PT or its diluent, or with a combination of the LTB4 receptor antagonist, LY-223982, and of the PAF receptor antagonist, BN 50730, and stimulated with increasing concentrations of A23187. These concentrations of the receptor antagonists were used on the basis of their being found to inhibit by ; 80% the elevation in free intracellular calcium elicited by either 30 nM LTB4 or PAF (data not shown). Cells exposed to A23187 alone required a minimal dose of 100 nM to allow the detection of 5-LO products derived from endogenous AA, with a maximal response being achieved at between 1 and 3 jAM A23187 (Figure 1). In cells pretreated with either PT or with the LTB4 and PAF receptor antagonists however, the formation of all AA-derived 5-LO products (i.e. LTB4, its w-oxidation products, and the LTA4 non-enzymatic degradation products) was markedly inhibited in the 100-300 nm range, the inhibition being less evident at A23187 concentrations which induced a maximal response. A parallel effect was consistently observed between the inhibition by PT and that by both receptor antagonists, suggesting that newly synthesized PAF and/or LTB4 contribute to ongoing LT synthesis induced by submaximal concentrations of A23187, and that they account for most of the G protein-mediated effects of A23187 on LT synthesis. To determine the respective contributions of endogenous PAF and LTB4 to LT synthesis in ionophore-challenged PMN, we investigated the effect of the PAF receptor antagonists, BN 50730 and L-659,989, and of the LTB4 receptor antagonist, LY-223982, either individually or in combination, as well as the effect of PT. At concentrations of A23187 which elicit a nearly maximal LT synthesis (i.e. 1.0 JiM), the receptor antagonists had a moderate inhibitory effect on LT synthesis (Figure 2a). In contrast, when a submaximal dose of A23187 was used (i.e. 200 nM), pretreatment with individual receptor antagonists yielded a profound inhibition of LT synthesis (Figure 2b). The combination of BN 50730 and LY-223982, as well as PT pretreatment, consistently led to a somewhat greater inhibition of LT synthesis.
A23187 300 nM
F-
Table 1 Effect of pertussis toxin or of PAF and LTB4 receptor antagonists on A23 187-elicited 5-LO activation under conditions where neither PAF nor LTs are synthesized in detectable amounts
0
U i: I-o
4w
.50
*
T,
BN 50730 lOOnM L-659,989 100 nM LY-223982 1.0 jiM BN + LY PT 0.5 jig ml- I
,_..A
-L Im
5j
0
.-I
_
Con BN L
LY BN PT
LY
Con
Figure 2 Effect of pertussis toxin, and of PAF and LTB4 receptor antagonists, on A23187-induced LT synthesis in human PMN. PMN were stimulated (5 min, 37C) with 1.0 jiM (a) or 300 nM (b) A23187, in the presence or absence (5 min before stimulation) of 1.0 jiM LY-223982 (LY), I00 nM BN 50730 (BN) or I00 nM L-659,989 (L), or following pre-treatment with PT (0.5 jig ml-', 3 h at 37QC). Samples were analysed by r.p.-h.p.l.c., as described in Methods. Mean ± s.e.mean of averaged triplicate determinations from at least 5 independent experiments. The 100% control (Con) values, representing the sum of LTB4 and of its w-oxidation products, averaged 162 ± 10 ng per 5 x 106 PMN (a, n = 7) and 123 ± 8 ng per 5 x 106 PMN (b, n = 8). *P< 0.04; **P< 0.01. For abbreviations, see text.
Compound
5-LO activation (% of control)
Number of experiments
p
104.2 ± 6.6 103.7 ± 4.1 98.2 ± 4.3 101.7 ± 9.1 92.7 ± 9.4
3 3 5 5 7
0.296 0.234 0.347 0.429 0.233
PMN suspensions were pretreated with the receptor antagonists (5 min, 37C), or with PT (3 h, 37C) and subsequently incubated for 10 min at 37°C with 3.0 jiM 15-HpETE, either alone or with 30 nM A23187. The conversion of 15-HpETE into 5,15-DiHETE (reflecting 5-LO activation) was then determined by r.p.-h.p.l.c. analysis, as described in Methods. Comparisons of 5-LO activity (expressed as % of control) have been calculated after subtracting the 5,15-DiHETE levels resulting from exposure of PMN to 15-HpETE only (± antagonists or PT). Values are the mean ± s.e.mean from the indicated number of experiments, each performed in triplicate. The 100% control values averaged 21 ± 1 ng 5,15-DiHETE per 5 x 10' PMN for A23187 30 nM (n = 7). BN + LY, simultaneous exposure to 100 nM BN 50730 and 1.0 jiM LY-223982. For abbreviations, see text.
AUTOCRINE EFFECTS OF LTB4 AND PAF IN NEUTROPHILS
To rule out possible nonspecific effects of the antagonists or of PT on 5-LO activity, we investigated whether these compounds would affect 5-LO activation induced by a low concentration (30 nM) of A23187, which allows 5-LO activation without triggering a detectable LTB4 or PAF synthesis. For this purpose, we used exogenous 15-HpETE, which is converted into 5(S),l5(S)-dihydroxy-6,8,l1,13 (E,Z,Z,E)eicosatetraenoic acid (5,15-DiHETE) by the 5-LO. We have shown previously that this approach allows the determination of 5-LO activation independently of endogenous substrate release (McDonald et al., 1991). PMN were therefore incubated with 15-HpETE and/or 30 nM A23187, following treatment with the receptor antagonists or PT (Table 1). A23187 alone did not elicit the formation of 5-LO products or of PAF in detectable amounts (data not shown). Under all conditions tested, no significant inhibition of the A23187induced transformation of 15-HpETE into 5,15-DiHETE (reflecting 5-LO activation) was observed. Similarly, the 5,15DiHETE synthesis resulting from exposure of the cells to 15-HpETE alone was unaffected by prior treatment with the antagonists or with PT (data not shown).
-
c
10o10o
0
A23187 1.0 gM
a
A23187 300nm
I
1r
MITM
Effect of endogenous LTB4 and PAF on A23187-induced 5-LO activation and AA release Our previous observation, that both PAF and LTB4 are potent 5-LO activators (McDonald et al., 1991; 1992), led us to determine whether the amplification of A23187-induced LT synthesis by endogenous LTB4 and PAF could take place at the 5-LO level. PMN were pretreated with PT or the receptor antagonists (individually or in combination), and then stimulated with 300nM or 1.0JIM A23187, along with 15-HpETE. Figure 3a shows that regardless of the ionophore concentration used, PAF and/or LTB4 receptor antagonists had no significant effect on A23187-induced conversion of 15-HpETE into 5,15-DiHETE in all but two cases, where a minor inhibition was noted, whilst PT had a moderate inhibitory effect at 300 nM A23187. We next examined whether A23187-induced AA release would be affected upon pretreatment of the cells with the receptor antagonists or PT. In these experiments, PMN were stimulated with either 1.0 JAM or 300 nM A23187. At the higher ionophore concentration, the PAF receptor antagonists failed to alter AA release significantly, whereas LY-223982 and PT exerted a moderate inhibition (Figure 3b, left panel). In contrast, all compounds potently inhibited AA release induced by 300 nM A23187 (Figure 3b, right panel).
Autocrine effects of endogenous LTB4 and PAF on LT synthesis induced by chemoattractants in rhGM-CSFprimed PMN Having shown that endogenous LTB4 and PAF enhance
1-11 0)
c:c
855
A23187-induced LT synthesis in an autocrine fashion, we determined to what extent these neutrophil products may contribute to LT synthesis elicited under physiologicallyrelevant conditions. PMN were pretreated with rhGM-CSF or its diluent, and during the last 5 min of pre-incubation, the cells were exposed to the PAF and LTB4 receptor antagonists, individually or in combination. Cells were then stimulated with either fMLP or rhC5a, and LT synthesis was measured. The effect of PT pretreatment was not investigated in these experiments, as it would have prevented PMN
50
w
w Ix V-
fri
b
1001f-I 1
-
0
T
0
UM
a
fMLP 300 nM
b
C5a100nM
100
50 0 C
a1)
C, 0
0N
C.
0
*
Con BN L
~~~~~~~~~~~~~~~~~~~1:1
LY BN PT Con BN L LY
LY BN PT LY
50 (A -j
-4
Figure 3 Effect of pertussis toxin, and of PAF and LTB4 receptor antagonists, on A23187-induced 5-LO activation and AA release in human PMN. Incubation conditions are as in Figure 2. (a) Ionophore-induced 5-LO activation (as determined by the 5-LOmediated conversion of exogenous 15-HpETE into 5,15-DiHETE). Samples were analysed by r.p.-h.p.l.c., as described in Methods. The given percentages have been calculated after subtraction of the 5,15DiHETE synthesis resulting from exposures of PMN to 15-HpETE only ( ± receptor antagonists or PT). Mean ± s.e.mean of averaged triplicate determinations from at least 3 independent experiments. The 100% control (Con) values averaged 512 ± 33 ng 5,15-DiHETE per 5 x 106 PMN (left panel, n = 3) and 270 ± 31 ng per 5 x 106 PMN (right panel, n = 5). (b) Ionophore-induced AA release (2.5 min stimulations). Samples were analysed by l.c.-m.s., as described in Methods. Mean ± s.e.mean of averaged triplicate determinations from 5 independent experiments. The 100% control values averaged 203 ± 27 ng AA per 5 x 106 PMN (left panel, n = 5) and 94 ± 3 ng per 5 x 106 PMN (right panel, n = 5). BN, BN 50730; L, L-659,989; LY, LY-223982. )P
